Talel Taieb
A
A
Data Analyst & Engineer – I build automated data pipelines and business dashboards.
From Airbus to freelance projects, I turn raw data into smart, scalable systems that improve visibility, decision-making, and performance.
Specialized in dbt, Airflow, Streamlit, and Python – let’s automate what matters.
About
Experience
- Built an end-to-end KPI automation pipeline using dbt, SQL, and Skywise to feed interactive dashboards for HR and product teams.
- Deployed real-time dashboards in Streamlit, Tableau, and Power BI — used by 5 cross-functional teams to track operations and priorities.
- Developed alerting systems using anomaly detection models (Isolation Forest + rule-based logic) to flag high-risk engineering changes.
- Extracted insights from technical documents with BERT-based NLP models, improving auditability and decision speed.
- Collaborated with engineers, HR leaders, and PMs to define KPIs, improve data pipelines, and ensure delivery of actionable tools.
- Tools: Python, SQL, dbt, Airflow, Streamlit, Tableau, Power BI, Skywise, scikit-learn, Transformers
Projects
A data dashboard for project managers, built with real-world e-commerce data.
- Tools: Python, Pandas, Plotly, Streamlit, Streamlit Cloud, Kaggle
- Built an interactive dashboard to track project performance, delays, and customer satisfaction.
- Integrated real business data from the Brazilian E-Commerce (Olist) dataset.
- Implemented dynamic filters (city, score, delay) and downloadable risk reports.
- Visualized KPIs, delay analysis, and customer sentiment in real-time with Plotly.
A full-stack ML app that predicts telecom customer churn with live inputs and risk insights.
- Tools: Python, Streamlit, scikit-learn, XGBoost, SHAP, SMOTE
- Built a live prediction web app with a 6-feature input form and user-friendly UI.
- Used a stacked ensemble (Logistic Regression + XGBoost + Random Forest) for high accuracy.
- Applied SMOTE to address class imbalance and enhance generalization.
- Enhanced interpretability using SHAP values and engineered informative features.
A multilingual NLP app that classifies customer reviews as positive or negative using BERT.
- Tools: HuggingFace Transformers, BERT, Deep Translator, Streamlit
- Built an NLP pipeline that auto-translates and classifies reviews in any language.
- Fine-tuned BERT model for high-accuracy sentiment classification.
- Interactive web app with real-time predictions and confidence scoring.
- Modern Streamlit UI and multilingual text support for global scalability.
Skills
Languages & Databases
Python
SQL
PostgreSQL
Bash
Snowflake
PySpark
Redshift
Python Libraries
Pandas
NumPy
Matplotlib
Seaborn
scikit-learn
DuckDB
Hugging Face
Frameworks & Dashboards
Streamlit
FastAPI
Tableau
Looker Studio
Power BI
Looker
Machine Learning & AI
TensorFlow
PyTorch
Spark MLlib
MLflow
Cloud & Tools
Git
Docker
AWS
Airflow
dbt
Education
Toulouse, France
Degrees: Double Master’s Degrees in Applied Mathematics & Computer Science
Specialization: Hybrid Artificial Intelligence (ML + Deterministic & Probabilistic Modeling)
Type: Apprenticeship-based Engineering Program (Airbus)
- Machine Learning & Deep Learning
- Probabilistic Models & Inference
- Optimization & Decision Systems
- Big Data Architecture
- Advanced Python & Data Engineering
Relevant Courseworks:
French Scientific Baccalauréat
Tunis, Tunisia
Specialization: Mathematics & Physics
Status: Independent Candidate (Candidat libre)
- Advanced Mathematics
- Physics & Chemistry
- Earth and Life Sciences (SVT)
- Philosophy
- French Language and Literature
- English as a Foreign Language
Core Subjects: